Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generation

Dimensions overall score 7.0

• A new distillation method for real-time video generation that improves quality and reduces inference costs by ensuring self-consistency across denoising steps and optimizing for KV cache usage. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion.problem 35%
  Evidencepartial

  A new distillation method for real-time video generation that improves quality and reduces inference costs by ensuring self-consistency across denoising steps and optimizing for KV cache usage. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion.

  Implicationpartial

  Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.

  Verificationpartial

  partial

• Distilling video generation models to extremely low inference budgets (e.g., 2--4 NFEs) is crucial for real-time deployment, yet remains challenging. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion.method 35%
  Evidencepartial

  Distilling video generation models to extremely low inference budgets (e.g., 2--4 NFEs) is crucial for real-time deployment, yet remains challenging. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion.

  Implicationpartial

  Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.

  Verificationpartial

  partial

• ScienceToStartup currently rates this 7.0/10 on the public viability pass. Across extensive experiments on both non-autoregressive backbones (e.g., Wan~2.1) and autoregressive real-time paradigms (e.g., Self Forcing), our method, dubbed \textbf{Salt}, consistently improves low-NFE video generation quality while remaining compatible with diverse KV-cache memory mechanisms. A public repository is linked, so build verification can inspect implementation evidence instead of treating the paper as PDF-only.result 35%
  Evidencepartial

  ScienceToStartup currently rates this 7.0/10 on the public viability pass. Across extensive experiments on both non-autoregressive backbones (e.g., Wan~2.1) and autoregressive real-time paradigms (e.g., Self Forcing), our method, dubbed \textbf{Salt}, consistently improves low-NFE video generation quality while remaining compatible with diverse KV-cache memory mechanisms. A public repository is linked, so build verification can inspect implementation evidence instead of treating the paper as PDF-only.

  Implicationpartial

  Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.

  Verificationpartial

  partial

• Generative Video moved forward this cycle; last verified April 2026. Public score 7.0/10. Implementation evidence is present through a linked repository.why now 35%
  Evidencepartial

  Generative Video moved forward this cycle; last verified April 2026. Public score 7.0/10. Implementation evidence is present through a linked repository.

  Implicationpartial

  Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.

  Verificationpartial

  partial

Competitive landscape

A new distillation method for real-time video generation that improves quality and reduces inference costs by ensuring self-consistency across denoising steps and optimizing for KV cache usage.